Implementing artificial neural networks through bionic construction.

Hu He,Zhiheng Xu,Guo Liu,Jinfeng Zhao,Yingjie Shang,Xu Yang,Ning Deng,Wenhao Zheng,Mengyao Ji,Liya Dong,Jie Zhang

doi:10.1371/journal.pone.0212368

Abstract

It is evident through biology research that, biological neural network could be implemented through two means: by congenital heredity, or by posteriority learning. However, traditionally, artificial neural network, especially the Deep learning Neural Networks (DNNs) are implemented only through exhaustive training and learning. Fixed structure is built, and then parameters are trained through huge amount of data. In this way, there are a lot of redundancies in the implemented artificial neural network. This redundancy not only requires more effort to train the network, but also costs more computing resources when used. In this paper, we proposed a bionic way to implement artificial neural network through construction rather than training and learning. The hierarchy of the neural network is designed according to analysis of the required functionality, and then module design is carried out to form each hierarchy. We choose the Drosophila’s visual neural network as a test case to verify our method’s validation. The results show that the bionic artificial neural network built through our method could work as a bionic compound eye, which can achieve the detection of the object and their movement, and the results are better on some properties, compared with the Drosophila’s biological compound eyes.

Highlights

Researches about artificial intelligence have become very popular in current days, due to the ever growing demands from application domains such as pattern recognition, image segmentation, intelligent video analytics, and autonomous robotics [1,2,3,4]
It is evident through biology research that, biological neural network could be implemented through two means: by congenital heredity, or by posteriority learning
We presented a method to implement artificial neural network through bionic construction rather through huge scale’s training and learning

Summary

Introduction

Researches about artificial intelligence have become very popular in current days, due to the ever growing demands from application domains such as pattern recognition, image segmentation, intelligent video analytics, and autonomous robotics [1,2,3,4]. 2. demanding a huge amount of computing power when trained with a large scale of data. Huge amount of computing power is needed to accomplish the training process This will become a bottleneck when the scale of data becomes even larger. It is unfeasible to accomplish intuition or reasoning using DNN This is mainly due to the fact that the traditional ANN, including DNN, uses essentially the same and fixed structure even for different problems. This can lead to a large number of redundancy in the network: 1. Training and learning process could implement powerful artificial neural network. The following is organized as: Section 2 discusses related work; Section 3 presents our method; Section 4 discusses the structure of Drosophila’s visual system briefly; Section 5 gives details of our test case research and the experiment result; we give the conclusion

Related work

Conclusion